welder orbital welding

4

Click here to load reader

Upload: seramon

Post on 14-Oct-2014

115 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Welder Orbital Welding

ACCORDING TO A WELL-KNOWN mechanical contractor whospecializes in the construction of bioprocessing systemsused for the production of pharmaceutical products, the

documentation required to support these installations can easi-ly account for 30% of the total construction labor hours on aproject. A significant number of these hours are spent tracking,accumulating and compiling all of the supporting data requiredto verify that every welded connection made on the tubing sys-tem meets the stringent specifications for hygienic applications.

Orbital tube welding, which is the recommended method offabrication for bioprocessing equipment and tubing systems,has not changed much since the 1960s, when it was introducedinto the aerospace industry. We still use the Gas Tungsten ArcWelding (GTAW) process and orbit or rotate an electrodearound a fixtured weld joint with a predetermined set of weldparameters programmed into the power supply.1

However, today’s orbital welding equipment utilizes micro-processor technology and has the ability to monitor the weldprocess, capture and compile real time data and electronicallytransfer it to a computer where it can be formatted for statisticalanalysis. Weld programs can now be automatically created andadjusted by the welding power supply. Weld logs can be main-tained electronically and imported directly to a computer wherethey can be easily formatted for installation qualification.

So although we may still perform an orbital tube weld muchthe same way we did 30 years ago, today we have much greatercontrol over the welding process and can obtain significantlymore data to verify it. Most significantly, the ability to transferdata to the computer for statistical analysis has the potential tochange the way industry approaches orbital welding.

New Weld Documentation MethodsWeld documentation is required by the FDA to provide fulltraceability for all materials and procedures used for construc-tion of bioprocessing systems. The documentation is required toprovide the necessary verification that each weld meets theacceptance criteria as defined by the weld specification. In theevent there is a problem with a system, such as contamination,the documentation provides a means to go back and revieweach welded connection as part of the effort to track down andfind the source.

The amount of documentation a contractor must submit tosupport a single weld is staggering, but of particular interest forthis discussion is the weld log kept by the welding operator.Throughout the construction project, orbital welding operatorsmust keep a daily weld log. Whenever a weld is completed, theoperator must manually record the event in the log. The weld isthen inspected, typically with a video boroscope, and then

properly identified with a unique identification number thattraces it to the weld log. At the end of the day, weld logs are col-lected from all of the operators and submitted to a quality con-trol administrator. The quality control administrator inputs thedata from all of the operators into a spreadsheet on a computerand prints it out. The weld data report is now in the properinstallation qualification format and is ready for submittal. Thecontractor is also required to create isometric drawings, depict-ing the completed welds with the corresponding identificationnumbers, and submit them with the installation qualificationreport.

Collecting and submitting all this documentation providesthe verification needed to ensure the integrity of each weldedconnection. Or does it? The weld log that is kept manually bythe operator is probably the weakest link in the weld documen-tation process because it is susceptible to human error. Theintegrity of the weld log is dependent on the operator remem-bering to record the appropriate information after each weld.We assume the information is recorded correctly (in the correctcolumn) and we hope the operator’s penmanship is readable bythe quality control administrator who must manually type theinformation into a computer.

Accumulating, compiling, and submitting weld documenta-tion drives up the individual cost per weld. For example, a jobwith 10 operators completing 20 welds each per day, six days aweek, would expect to complete a total of 1,200 welds per week.Using a conservative estimate of six minutes total for the com-bined time of logging and inputting each weld would result in7,200 minutes per week or 120 hours. At $50 per hour, the costfor manually documenting welds would amount to $6,000 perweek. This figure does not include the additional costs of gen-erating other documentation, such as slope reports, bills ofmaterials and weld map drawings. A typical cost per weld for ahigh purity system without this detailed level of documentationis approximately $80 to $100. Therefore, the same weld includ-ing manual documentation can easily cost as much as $150. Ifthere is an error in the documentation, the weld will be suspect;in the worst case, the weld would have to be cut out andreplaced at the contractor’s expense.

Today, the weld log can be kept electronically, thanks toadvances in microprocessor controlled orbital welding powersupply technology. The operator no longer has to record anyinformation. Upon the completion of each weld, the data need-ed for the weld log is stored in memory on board the power sup-ply. At the end of the day, the weld log data is still collected andsubmitted to the quality control administrator, but now it is in

ORBITAL WELDING

Orbital WeldingNew documentation methods simplify validation of stainless steel systems

This article first appeared in the March 2001 issue of CONTRACT PHARMA • 1

By William J.WuertzSwagelok

William J.Wuertz is manager, welding systems, at Swagelok.

Photo courtesy of Swagelok

Page 2: Welder Orbital Welding

2 • This article first appeared in the March 2001 issue of CONTRACT PHARMA

ORBITAL WELDING

the form of a PC data card. The card is read by the computer andthe data is imported into a common spreadsheet or databasewhere it is formatted for the installation qualification report. Theelectronic weld log is completed, requiring much less time andeffort on the part of the operator and the quality control admin-istrator, at a significant cost savings for the contractor.

Maintaining the weld log electronically requires a little moretime up front to input setup data into the power supply, but

once this is done, usually only periodic updates are needed.User interface with the welding power supply has alsoimproved. Modern keypads and crisp LCD screens have beenintegrated into the machines, making the task of data inputmuch easier. Power supply operating systems have beendesigned to look and perform in a manner similar to those usedwith personal computers, so they are familiar to the user.Navigating the software has been made easy through program-ming that is intuitive, logical and easy to master.

The amount of information available from an electronic weldlog is significantly greater than what is kept on a manual log.The additional information is a valuable asset to the traceabili-ty documentation. Following is a list of the some of the infor-mation that may be found on an electronic weld log:

• Date and time • Weld count number • Power supply model number• Power supply serial number• Procedure identification• Operator identification• Programmer identification• Project identification• Drawing identification• Weld joint description • Weld joint material type• Material heats• Tube diameter• Tube wall thickness• Shielding and purge gas identification numbers• Flow rates for shielding and purge gas• Weld head serial number• Electrode identification• Software revision level• Date of last calibration• NotepadThe dimensional data usually can be configured to be pre-

sented in metric or fractional units. Most power supplies willeven convert logged data between fractional and metric. Inaddition to this input data, the output values of travel speed,time, average current, voltage and downslope will be included

in the weld log. Some machines even calculate and include kilo-joules per inch, which defines the heat input.

Benefits of Real Time DataOrbital welding equipment today is capable of acquiring realtime data output and exporting it to a computer where it can beformatted for verification and analysis. The acquisition of realtime data simply means capturing and storing the output val-ues of the welding parameters while the weld is in progress.The welding parameters that are being tracked are those criticalto the welding process; these include, but are not limited to, thetravel speed (rotation speed of the electrode around the weldjoint), the average current and the voltage. Utilizing the powerof the microprocessor, the power supply samples these param-eters at approximately 50 times per second and records theresults in memory with the weld log.

Much can be determined by analyzing the weld parameters.For example, variations in the voltage may indicate a change inthe arc gap (distance between the electrode and weld joint),which could be the result of ovality in the weld joint. It couldalso indicate that the weld joint was not properly prepared andthat there is a gap or space present in the weld joint. Variationsin the current may be the result of a large spike in the incomingpower source to the power supply. There are numerous otherpossible conclusions that could be made regarding differentdata variations. Without the data, such an analysis would notbe possible.

Statistical Process Control (SPC)Statistical process control (SPC) is defined as “the application ofstatistical techniques for measuring and analyzing the variationin processes.”2 The acquisition of real time weld data providesthe opportunity to apply SPC to orbital welding and improvethe quality of the process. Realtime data provides the feedbackloop needed to better understand how to control the processand optimize the output. By analyzing live data, or measuredperformance, we can determine process quality characteristicssuch as travel speed, average current, arc gap, weld joint posi-tion, shielding and backing gas, purge pressure and materialproperties. With enough data, we can establish acceptable per-formance standards. The control of the performance can then bemeasured through the feedback loop or realtime data and com-pared to the standard; any difference can be acted upon.

Orbital welding equipment suppliers have understood thevalue of SPC and now provide welding equipment outfittedwith the tools necessary to implement it. The power suppliesare capable of being programmed with preset acceptable stan-dard deviations for the critical output parameters, so that, as theoutput data is being sampled, it is also being checked againstpreset acceptable limits. Should the output exceed the accept-able limits, an alarm can be enabled to sound and alert the oper-ator. When this occurs, the printout provided by the power sup-ply includes a statement that the performance was unaccept-able. The actual output data appears on the printout, so theoperator or quality control engineer is able to review it in aneffort to determine what happened. The electronic data sent tomemory also includes the actual data and a statement that theperformance was unacceptable.

. . . With enough data, we can establish acceptableperformance standards.The control of the perform-ance can then be measured through the feedbackloop or realtime data and compared to the stan-dard; any difference can be acted upon . . .

Page 3: Welder Orbital Welding

ORBITAL WELDING

Power supplies are also capable of setup diagnostics, welddiagnostics, and machine diagnostics. The diagnostics take theform of messages that are normally displayed on the front paneldisplay screen. In addition, the diagnostics are recorded on theprintout from the power supply as well as stored in memory withthe weld log data. The setup diagnostics typically include opera-tor messages, such as invalid weld procedure, printer paperempty, PC data card not installed, etc. Weld diagnostics are mes-sages that describe the cause of a functional problem should oneoccur. Examples of these messages would include: rotor jam, mis-fire, arc out, and tasks not complete. Finally, the machine diag-nostics provide messages about the power supply, such as over-heating, minimum voltage alert, etc. Equipment today can recordnearly every event and make it available for review.

SPC can also be applied to the weld procedures used in theorbital welding process to produce a more uniform weld qual-ity. Weld procedures often include detailed setup specifica-tions, which play a key role in the control process for SPC. Forexample, the arc gap (which is the distance between the tip ofthe electrode and the work piece) is a parameter that plays asignificant role in the orbital welding process. Even a differ-ence of 0.005 inch can result in a variation of a completedweld. Because the arc gap is a critical parameter, the operatormust, upon initial setup, install the electrode in the weld headand position it, to the best of his ability, to obtain the desiredarc gap. Throughout the course of the job, the electrode willdeteriorate and will need to be changed. How does the opera-tor know when to change the electrode? How can the operatorbe sure that the new electrode will be installed in the sameposition the original electrode was for initial setup andcoupon testing?

SPC may provide the answer to these questions. Analyzingweld data for repeated welds using the same electrode mightreveal a change in the process that can be attributed to electrodedeterioration. If this can be established, an optimum number ofwelds per electrode might be determined and become part ofthe weld procedure.

There is a simple solution that will improve the likelihood thatthe arc gap is set in the right position during setup and that it isset in the same position every time it is changed. Equipmentsuppliers provide simple gages or procedures that are used toset the position of the electrode. The gages are set to a specificposition and then locked in place. Whenever the electrode ischanged, the gage is used to reposition the electrode to exactlythe same place it was on the initial setup.

Similarly, the weld joint must be positioned in the fixture sothat it is aligned with the electrode. A shift of just a few thou-sandths of an inch in the joint position can result in a weld withincomplete penetration. Simple gage blocks supplied with thefixtures provide a means to position the weld joint in the centerof the fixture every time. Including the use of such gages withthe weld procedures can improve setup protocol and result in amore consistent weld process.

Industry’s knowledge of material chemistry and understand-ing of its affects on weldability has significantly improved overthe last ten, or even five years. Today, material suppliers areproducing with controlled chemis-tries, specifically designedfor improved weldability and corrosion resistance. As a result,

material is becoming less of a variable for welding. The morewe understand about the relationship between welding andmaterial chemistry, the closer we become to being able to pre-dict the exact weld parameters needed to perform a weld.

Currently, corrosion testing of stainless steel welded jointsseems to be an ongoing effort in different industries for differ-ent reasons. The testing normally includes material of knownchemistries, input data defining the weld program and calcu-lations that describe the heat input. The corrosion tests provide

documented results and physical samples. But what is oftenmissing from the documentation is the realtime weld data. Theability to capture and analyze realtime data can help us gain abetter understanding of material chemistry and welding.

Although our understanding of material weldability hasevolved considerably in recent years, there is still a great dealwe do not know. Using real time data to analyze weldabilityissues, such as sulfur content, will be a valuable tool in improv-ing methods or finding solutions. The analysis of weld datamay indeed someday provide us with the ability to predict theexact weld parameters needed to perform a weld based on achemical analysis of the material.

Predicting weld parameters is being done right now withvery impressive results. Some orbital welding power suppliesnow have programs built into the software that will automati-cally create a program based on the material chemistry.Realtime data analysis, combined with existing weld programdevelopment practices, has enabled these programs to bedeveloped. Creating a new program is now a matter of goinginto the programming mode on the power supply andresponding to prompts from pull-down lists for weld configu-rations and materials. Then, with the simple push of a button,the power supply will execute a series of calculations andinstantly generate and load a weld program. If, after trying theweld program, it is determined that an adjustment is necessary,the user is again prompted by the power supply to select thesegment or location on the tube where a change in heat inputand subsequent weld penetration is desired. After making theselection, the machine can instantly create a new program witha small adjustment made to the heat input in the selected loca-tion. Manually determining weld program parameters requireshand calculations, reference tables and manual data input,which can be a difficult and time-consuming job, and oftenrequires a highly skilled programmer or operator.

The ability to generate weld programs quickly can bring animmediate benefit to verifying the weldability of incomingmaterial. Often, as tubing and other components are received,they are checked in, the paperwork is verified and the material

This article first appeared in the March 2001 issue of CONTRACT PHARMA • 3

ORBITAL WELDING

. . .The 1960s brought orbital welding to industry.In the 1980s, we began to learn how importantmaterial chemistry is to weldability and orbitalwelding. The 1990s may be remembered as thedecade of data . . .

Page 4: Welder Orbital Welding

4 • This article first appeared in the March 2001 issue of CONTRACT PHARMA

ORBITAL WELDING

receiving report is filed. If the material has a weldability issue,it may not be discovered until it is pulled from inventory andtaken to the work site. At this point, the material may have to bereturned and replacement material must be ordered and expe-dited in. If weldability test samples are made as material isreceived, this situation may be avoided. The programs that arecreated to perform the weldability tests on material as it isreceived can be saved until they are needed later when thematerial is being used.

Statistical Quality Control (SQC)Statistical quality control is defined as “The application of sta-tistical techniques for measuring and improving the quality ofprocesses.”3 Statistical analysis of real time data can be appliedto quality control practices currently being used with orbitalwelding. The power of the computer can be used to scan dataand look for irregularities. Spotting an unusual number amonghundreds of other numbers may be a challenge for the humaneye, but it will appear as a flashing yellow light to a computer.

A bit of data outside an acceptable window of deviationmay be a clue to a bigger problem. The opportunity now existsto re-inspect the suspect weld and the supporting data to veri-fy its acceptability. As an example, consider that a constructionjob with 1000 welds may have had 5 welds rejected. Is theresomething that the 5 rejects have in common that may also becommon with other welds in the system? Were they all weldedby the same operator? Did they all involve the same material?Were they all welded within the same hour? Computer analy-sis of the data provides a practical and timely way to sortthrough the weld data and look for such commonaltiesbetween the rejected welds. If a particular link is found, asearch through all the other weld data can be easily conductedto look for other possible suspects.

We know that variations in material properties can affectweldability. How material is selected and where it comes fromcan make a significant difference in welding productivity. Dataanalysis can provide information that may be useful when spec-ifying tubing, fittings and other components. For example, fit-tings with poor control over the dimensional characteristics canyield irregular weld beads due to ovality or wall thickness vari-ations. Tubing or other components that frequently presentweldability problems may be supplied from source that shouldbe evaluated. Data analysis may also provide the necessaryinformation to allow users to match tubing and components ofsimilar material chemistries in an effort to improve the likeli-hood of compatibility for welding.

Realtime Data Reports Improve ManagementCompiling detailed weld data into a database makes it possibleto generate different reports for different purposes. A databaseis an invaluable tool for managing activities such as productiv-ity analysis, costing, expense management and quality control.For example, a project manager preparing a proposal for a con-struction job may be looking for costing information on weld-ing. A query of the database could quickly provide a wealth ofinformation regarding the average costs associated with weld-ing different sizes, configurations and materials. This kind ofactual costing information would also be very useful to justify

bids based on time and material.Welding productivity is certainly an important issue that

must be managed properly to be competitive and profitable.Productivity statistics, such as weld per hour and welds per dayfor each operator, provide the visibility managers need to estab-lish a benchmark for their productivity. Further data analysiscould result in determining key initiatives to improve produc-tivity. One common productivity improvement is training; theeffectiveness of a training program effort can easily be measuredthrough routine reports that measure productivity. The reportsmay also indicate that equipment is the bottleneck and, thereby,provide justification for the selection of new equipment.

The quality control engineer can also use data analysis toimprove the quality of the welding process. Investigation intothe primary causes for weld rejects allows the quality controlengineer to develop and implement actions to prevent or mini-mize them from reoccurring. The database could provide areport that lists the most frequently encountered quality prob-lems sorted by the number of occurrences. For example, if anissue such as misalignment or discoloration should appear onthe list, the weld fixturing, purge practices, or setup protocolshould be reviewed.

The 1960s brought orbital welding to industry. In the 1980s,we began to learn how important material chemistry is to weld-ability and orbital welding. The 1990s may be remembered asthe decade of data.

Advanced welding power supply technology has improvedweld data collection and analysis. Real-time data collection andelectronic weld logs make weld information readily available.Weld data analysis can provide information for:

• Documenting welds to provide full traceability.• Controlling the welding process and optimizing output.• Producing a more uniform weld quality.• Predicting weld parameters.• Creating weld programs.• Managing welding productivity.• Justifying bids.In the past, it was not possible to easily acquire the necessary

data and subject it to statistical review. Today, with easy accessto complete weld data comes the ability to analyze and imple-ment quality tools such as SPC and SQC. In addition, electron-ic weld logs will bring relief to contractors who must diligentlycompile what seems like mountains of documentation for weldverification.

Advancements in microprocessor-controlled orbital weldingpower supply technology offer an opportunity to improvewelding processes and procedures. Our challenge now is to notbe too busy doing things the old way to learn how to do them abetter way—a way that can change industry’s approach toorbital welding technology. ■

References1.An American National Standard: Bioprocessing Equipment, October17, 1997, The American Society of Mechanical Engineers, New York,1997.2. J.M. Juran and Frank M. Gryna, Juran’s Quality Control Handbook,4th ed., McGraw-Hill, Inc., New York, 1988, p.24.2.3. Ibid.